I need a help to find a counter example (or to prove) the following

Let $G$ be a compact abelian topological group. Assume there exists a dense subgroup $H\leq G$ such that every element of $H$ is of finite order then $G$ is a Lie group (equivalently, is it isomorphic to a direct product of a torus with a direct product of finite groups).

Note that the converse is true, every compact (abelian or non-abelian) Lie groups contains a dense subgroup of torsion elements

  • 2
    $\begingroup$ Hmm, your notion of a Lie group looks too wide for me. The subgroup $t(G)$ of all periodic elements is dense of any compact abelian topological group $G$ of the form $\Bbb T^\kappa\times \prod A_\alpha$, where $\{A_\alpha\}$ is a family of finite abelian group whose periods are bounded in common. Or you are asking whether such form of $G$ is necessary to have the subgroup $t(G)$ dense? $\endgroup$ – Alex Ravsky Jun 1 '17 at 3:42
  • $\begingroup$ yes exactly, i asked whether such form is necessary to have the subgroup t(G) dense. $\endgroup$ – Yanko Jun 1 '17 at 17:13
  • 1
    $\begingroup$ Just for general knowledge, if G is countably based profinite group (so no $S^1$ parts) and t(G) is dense in G then G is a direct product. $\endgroup$ – Yanko Jun 17 '17 at 22:41
  • $\begingroup$ @Yanko what do you mean by "is a direct product"? $\endgroup$ – YCor May 6 at 23:16

There is a lot of counterexamples, but some efforts are needed in order to prove that a candidate group is indeed a counterexample. Also I may ask my teacher about them. The first family consists of zero dimensional candidates, which are closed subgroups of the product $\prod A_\alpha$ of a family $\{A_\alpha\}$ of finite abelian group whose periods are bounded in common. Candidates of bigger dimensions constitute the second family is so-called exotic tori, introduced by Dikranjan and Prodanov, I guess in [DP], that is compact abelian groups $G$ such that torsion part $t(G)$ has-non zero intersection with each non-zero closed subgroup of $G$. For us is essential that by Proposition 3 from [DP], torsion part of an exotic torus is dense in it. From the other hand, the motivation for introducing the exotic tori in [DP] was the fact that they are precisely the completions of the torsion minimal abelian groups. The structure of exotic tori has been described in detail in [DP]. In particular, Proposition 2.5* describes the structure of $n$-dimensional exotic tori as follows. A compact abelian group $G$ of dimensional $n$ is an exotic torus if and only if there is a representation of $G$ as a projective limit $$G=\lim_{\longleftarrow} G_m$$ such that $G_0=\Bbb T^n$, the continuous homomorphisms $\sigma_m$ in the projective system $$G_0\stackrel{\sigma_1}\longleftarrow G_1\stackrel{\sigma_2}\longleftarrow G_2\cdots \stackrel{\sigma_m} \longleftarrow G_m\longleftarrow\cdots$$ are epimorphisms and $\ker\sigma_m $ is a compact $p_m$-group for each natural $m$. Thus, according to Example 1* of [DP], for each natural $n$ there is a continuum non-isomorphic connected $n$-dimensional exotic tori (so here we have continuum many connected counterexamples). For more details on exotic tore see [DP] or [DPS]. See also [CD], in particular, Theorem 4.1 and Remark 4.2(c). Also you can google for exotic torus Prodanov Dikranjan.


[CD] W. W. Comfort, Dikran Dikranjan, On the poset of totally dense subgroups of compact groups, Topology Proceedings, 24, Summer 1999, 103–127.

[DP] D. Dikranjan, I. Prodanov, A class of compact Abelian groups, Annuaire Univ. Sofia, Fac. Math. Méc. 70 (1975/76), 191–206.

[DPS] D. Dikranjan, I. Prodanov, L. Stoyanov, Topological groups. Characters, dualities and minimal group topologies, Marcel Dekker, Inc., New York, 1990.

  • $\begingroup$ I might have a use of the article of DP, if you still have it then it would be nice if you can scan it for me. $\endgroup$ – Yanko Jul 26 '17 at 15:03
  • $\begingroup$ @yanko OK, if this is urgent then I can do this. If not then I can assign this task to the student and I hope it will be done in a pair of weeks from now (such a long delay is because there are vacations now, and it is not so easy for me to catch the student). $\endgroup$ – Alex Ravsky Jul 27 '17 at 20:55
  • $\begingroup$ That's fine, I'll wait. Thanks $\endgroup$ – Yanko Jul 30 '17 at 16:28
  • $\begingroup$ @yanko Yesterday I gave the paper to the student. He said that he'll start to scan it next week. $\endgroup$ – Alex Ravsky Jul 31 '17 at 22:09
  • $\begingroup$ @yanko At last I received the link to the scanned paper (added to the references). $\endgroup$ – Alex Ravsky Oct 31 '17 at 8:11

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.